Budburst phenology and host use by brumata (Linnaeus, 1758) (: Geometridae) in three Mediterranean species Yaussra Mannai, Olfa Ezzine, Axel Hausmann, Said Nouira, Mohamed Lahbib Ben Jamâa

To cite this version:

Yaussra Mannai, Olfa Ezzine, Axel Hausmann, Said Nouira, Mohamed Lahbib Ben Jamâa. Budburst phenology and host use by Operophtera brumata (Linnaeus, 1758) (Lepidoptera: Geometridae) in three Mediterranean oak species. Annals of Forest Science, Springer Nature (since 2011)/EDP Science (until 2010), 2017, 74 (1), pp.3. ￿10.1007/s13595-016-0600-3￿. ￿hal-01702516￿

HAL Id: hal-01702516 https://hal.archives-ouvertes.fr/hal-01702516 Submitted on 6 Feb 2018

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Annals of Forest Science (2017) 74: 3 DOI 10.1007/s13595-016-0600-3

ORIGINAL PAPER

Budburst phenology and host use by Operophtera brumata (Linnaeus, 1758) (Lepidoptera: Geometridae) in three Mediterranean oak species

Yaussra Mannai1,2 & Olfa Ezzine1,2 & Axel Hausmann3,4 & Said Nouira2 & Mohamed Lahbib Ben Jamâa1

Received: 12 June 2016 /Accepted: 28 November 2016 /Published online: 6 February 2017 # INRA and Springer-Verlag France 2017

Abstract weekly. Larval performance of O. brumata on the three oak & Key message Operophtera brumata L. performance species was analyzed. DNA extraction, PCR, and DNA se- varies among three Mediterranean oak species. Quercus quencing were performed. canariensis Willd is more susceptible to infestation proba- & Results Budburst of Q. canariensis and Q. afares was earlier bly due to its (i) early leafing, (ii) high nutritional value for than Q. suber. Q. canariensis was the most infested host. the larvae, and (iii) widespread abundance. Larvae which fed on Q. canariensis had faster development, & Context Larvae of Operophtera brumata were observed for lower mortality, and higher pupal weight than larvae fed on Q. the first time in an outbreak in Tunisia affecting Quercus afares and Q. suber. Molecular analyses showed that Tunisian canariensis, Quercus afares Pomel, and Quercus suber L. haplotypes were not different from those in Spain, , and Due to its polyphagous nature and the important ecological Germany. and economic damage it causes, it is most relevant to under- & Conclusion Results indicated differences in larval perfor- stand its interaction with North African species. mance. Q. canariensis was the most favorable host species. & Aims In this paper, budburst phenology of the three oak Its high density in the field and early leafing coinciding with species, larval performance, and genetic patterns of larval hatching made this species particularly susceptible. O. brumata were studied in northwestern Tunisia. & Methods In the spring of 2010, 2011, and 2012, budburst Keywords Q. canariensis . Q. afares . Q. suber . Winter phenology of host species and larval densities were monitored . Tunisia

Handling Editor: Aurelien SALLE Contribution of co-authors Yaussra MANNAI: project design, field and 1 Introduction laboratory data collection, data analysis, and paper writing Olfa EZZINE: project design, data analysis, and paper writing Budburst timing varies among and within tree species (Van Axel HAUSMANN: data analysis and paper writing Said NOUIRA and Mohamed Lahbib BEN JAMÂA: supervisors Dongen et al., 1997). Advances or delays in leafing are

* Yaussra Mannai Mohamed Lahbib Ben Jamâa [email protected] [email protected]

1 Laboratoire de Gestion et de Valorisation des Ressources Forestières, Olfa Ezzine National Institute for Research in Rural Engineering Water and [email protected] Forest (INRGREF), Bp 10, 2080 Ariana, Tunisia 2 Faculté des Sciences de Tunis, Campus Universitaire, El Manar, Axel Hausmann 1002 Tunis, Tunisia [email protected] 3 Bavarian State Collection of Zoology (ZSM), Munich, Germany Said Nouira 4 Staatliche Naturwissenschaftliche Sammlungen Bayerns (SNSB), [email protected] Munich, Germany 3 Page 2 of 8 Annals of Forest Science (2017) 74: 3 important for life cycles (Foster et al. 2013). The close of larvae among years and host plants, (ii) feeding be- coincidence of budburst and larval hatch of spring-feeding havior during budburst and larval performance, mea- generalist moth species was reported in many studies sured by larval development time, larval mortality, and (Hunter 1992; Tikkanen and Julkunen-Tiitto 2003; Van Asch pupal weight on the three host species. In addition, the et al. 2010;Fosteretal.2013). Such synchronization is im- DNA barcode fragment of the COI gene was sequenced portant for the survival and growth rate of the larvae, and for O. brumata collected from these three Quercus spe- therefore also, the expected fitness of the . For instance, cies and compared with DNA barcodes from another in order to gain maximal weight, the larvae of the , Tunisian population (Mzara forest), from various Operophtera brumata L., an important forest defoliator (Van European populations (Spain, Italy, and Germany) and Dongen et al. 1997), need to enter a bud at budburst to feed from individuals of Scharfenbe as upon the young leaves (Van Dongen et al. 1997). The dispers- outgroup to test the association between genetic pat- al behavior of O. brumata larvae might change with the onset terns and host use as well as test the hypothesis of a of budburst (Hunter 1990). The newly hatched larvae often do recent colonization of O. brumata in North Africa. not find suitable foliage on their natal tree and are forced to disperse. They can do this by “ballooning” (Holliday 1977; Hausmann and Viidalepp 2012). 2 Materials and methods O. brumata is a Holarctic species (Winstad et al. 2011; Hausmann and Viidalepp, 2012) widely distributed in 2.1 Study area Europe which has rapidly expanded its range colonizing other continents. It was recently reported as new for North The study site is located in the Ain Zena forest in northwestern Africa (Hausmann and Viidalepp 2012) and Tunisia Tunisia(alt.950m,36°43′ N, 8° 51′ E) at the southern edge of (Mannaietal.2015). Such extension of the distribution a large forest in Ain Draham. Vegetation is dominated by area offers the opportunity for the winter moth to contact Q. canariensis, Q. afares,andQ. suber (Mhamdietal. and use new host plant species. In Tunisia, the main host 2013). Q. afares is an endemic North African species origi- species is Q. canariensis Willd but Quercus afares Pomel nating from hybridization between Q. suber and and Quercus suber L. are also attacked (Mannai et al. Q. canariensis (Mir et al. 2006). The average height of the 2015). O. brumata is univoltine (Van Dongen et al. three species at the study area is 7.3, 12, and 18 m for Q. suber, 1997), but the time for egg and larval developments, the Q. afares,andQ. canariensis,respectively. length of pupation period and adult emergence varies throughout its range: in southern Italy, the egg stage lasts 2.2 Tree density about 2 months and pupation for 8 months (Horgan 1993), whereas, in northern Europe, the egg stage lasts about The density of Q. canariensis, Q. suber,andQ. afares was 8 months and pupation lasts for about 3 months (Horgan estimated by counting the number of trees in five 400 m2 1993). Depending on the weather, the needs 1– plots, totaling an area of 0.2 ha. 2 months from hatching to maturity passing through five instars (Kúti et al. 2011). Adults emerge in the autumn or 2.3 Budburst phenology and larvae density mid-winter, usually late October to early December in Central Europe, but this shifts to a late winter phenology In 2010, 2011, and 2012, samples were taken on a weekly (December to March) in southernmost Europe and North basis from mid-March to late April, for 6 weeks (W1–W6) Africa (Hausmann and Viidalepp 2012). to collect larvae and 9 weeks for budburst, until the first of Early field observations of O. brumata in Tunisia May (W1–W9). Every week, two branches from 10 mature suggested restricted patterns of host use with feeding trees per host species were monitored, one low-level branch activity concentrated on only three host plants of the (2.5 to 5 m) and one from crown height level (>5 m). Branches genus Quercus: Q. canariensis mixed with Q. suber were carefully cut using a pole pruner and bagged in a large and Q. afares in the Ain Zena reserve (Mannai et al. plastic bag to avoid losing larvae. In the laboratory, branches 2015). The use of novel hosts leads us to hypothesize were used to count larvae density and the proportion of swol- firstly that some host adaptation or specialization might len buds (phenological stage in which O. brumata can colo- occur at regional/species level. Secondly, budburst phe- nize buds (Hunter 1990)). nology may play an important role which affects the interaction between this polyphagous insect and the 2.4 Laboratory feeding trials host plant used by larvae. Field data and laboratory experiments were combined in order to investigate The performance of O. brumata on Q. canariensis, Q. afares, whether O. brumata showed differences in (i) density and Q. suber was compared in laboratory feeding trials. Annals of Forest Science (2017) 74: 3 Page 3 of 8 3

Experiments were performed in the spring of 2011, coinciding of trace files as well as exclusion of chimaera and sequences with the budburst of the host plants. Neonate larvae were with stop codons. DNA extracts are stored at the CCDB, with obtained from branches collected in the field in March 2011 aliquots being deposited in the DNA Bank facility of the ZSM and were individually placed in Petri dishes, kept at 25 ± 2 °C (see http://www.zsm.mwn.de/dnabank/). Sequences and and a light regime of 12:12 L:D (light:dark) as in natural metadataarehostedinBOLD(BarcodeofLifeDataSystems, conditions and reared ad libitum on leaves of each tested project INRGR “Global Geometridae/Lepidoptera of Tunisia- oak species. Thirty larvae, repeated 3 times, were used for cork oak defoliators-INRGREF”) and are accessible and down- each tested plant. Young still-expanding leaves were collected loadable in the public dataset DS-OPEROPH. All sequences are daily from plantlets of each species, planted in the same con- deposited also in GenBank according to the iBOL data release ditions at the nursery of the INRGREF. Larvae were checked policy. Sequence ID numbers on BOLD are provided in Table 1. daily and numbers of molted larvae were recorded. Larval Images, GPS coordinates, and sequence trace files for each development time for each tested oak species was reported. specimen as well as details on host institution can be obtained To evaluate larval performance, development time, larvae sur- from the Barcode of Life Data System (BOLD; Ratnasingham vival, and pupal weight of the 5th instar larvae were assessed and Hebert 2007), public DS-OPEROPH. on each tested species as biological parameters. Eight sequences of German specimens of O. brumata (A. Hausmann), one of a southern Italian specimen (M. Infusino) 2.5 Mortality and pupal weight of field collected larvae and one sequence of a southern Spanish specimen (A. Hausmann) were included into the analysis. German and First, second, and third instar larvae were collected only from Italian sequences of 8 individuals of O. fagata were used as Q. canariensis and Q. afares as they were absent on Q. suber. outgroup. A first analysis was performed with the tools of Fourth and fifth instar larvae were collected from Q. canariensis, BOLD database and was then refined on MEGA6 and Q. afares,andQ. suber. Each larva was kept individually in a Petri MEGA 7 (Tamura et al., 2013; Kumar et al. 2016) construct- dish and reared to pupation on young expanding leaves of each ing a Maximum Likelihood (ML) Tree including 21 host species collected daily from plantlets at the INRGREF nurs- Operophtera specimens from the western Palearctic (3 from ery. The Petri dishes were examined daily to record the number of Tunisia), bootstrap method, 500 replicates, Tamura-Nei mod- molting and dead larvae. Rearing tests were done in the spring of el, complete deletion, bootstrap values indicated when >50%. 2011. Pupal weight of the 5th instar larvae collected from the field Alignment was based on the alignment tools of BOLD data- and larvae developed in the laboratory (from the second instar to base. Manual alignment check revealed no errors (Fig. 4). pupation) on Q. canariensis, Q. afares,andQ. suber were com- pared in orderto investigate thesuitability of these hosts forwinter 2.7 Statistical analysis moth larvae. The statistical analysis was performed using the SPSS-10.0 2.6 Molecular analysis software package for Windows. Generalized linear models (GLMs) were applied to the fol- Mature larvae were collected by hand from the host plants lowing dependent variables: (1) the Julian day when 50% of (Q. canariensis, Q. afares,andQ. suber) in April 2011 from budburst occurred; (2) the number of larvae per branch, con- Ain Zena and Mzara. To prevent the sampling of siblings, each sidering tree species and year as factors; (3) larvae develop- larva was collected from a different tree. Pupae were collected ment (number of days spent in each instar), considering tree from the soil next to infested trees. Larvae and pupae were species as a factor. A Normal distribution model best fitted the preserved in 96% ethanol. One or two segments of larval thorax Julian day when 50% of budburst occurred and larvae devel- and the cremaster part of pupae were sampled into lysis plates opment. A Poisson distribution model best fitted the number for DNA barcoding. In total, 15 individuals were sampled. of larvae per branch. The effect of each tested oak species on DNA extraction, PCR, and DNA sequencing were performed the larval development time and the pupal weight was at the Canadian Centre for DNA Barcoding, Guelph, Canada assessed with an analysis of variance (ANOVA) and (CCDB), following standard high-throughput protocol, that can complemented by multiple comparisons of means by the be accessed under http://ccdb.ca/resources.php.PCR SNK test (Student–Newman–Keuls) and was expressed as amplification with a single pair of primers (Ivanova et al. 2006 mean ± standard error of mean (MSE). ) consistently recovered a 658-bp region near the 5′ terminus of The proportion of dead larvae among the total individuals the mitochondrial cytochrome c oxidase 1 (CO1) gene that in- obtained in the feeding experiments was analyzed by GLM cluded the standard 648 bp barcode region for the king- using a Binomial model with log link function, considering dom (Hebert et al. 2003). PCR primers used were LepF1/LepR1 the factor plant species. Results are presented in the form of (Hebertetal.2003). Quality check of the sequence data follow- the Wald’s chi-square test value (χ2), parameter estimates and ed the CCDB standards and included the accurate examination the respective P value. 3 Page 4 of 8 Annals of Forest Science (2017) 74: 3

Table 1 Insects used for molecular analyses: 21 specimens Species Stage Sequence-ID number Location Host plant of Operophtera brumata and in BOLD O. fagata. BOLD sequence ID- numbers, sites, and host plant Operophtera brumata GWOSP600-11 Ain zena (Tunisia) Q. afares GWOSP595-11 Mzara (Tunisia) Q. canariensis GWOSP587-11 Ain zena (Tunisia) Q. canariensis Adult GBLAC679-13 Bavaria, Oberbayern (Germany) – GBLAC990-13 Bavaria, Oberbayern (Germany) – GBLAC166-13 Bavaria, Oberbayern (Germany) – GWOSP887-11 Sicily (Italy) – GWOTD346-12 Andalusia (Spain) – GBLAA454-14 Schleswig-Holstein (Germany) – Larva GWORB1495-08 Bavaria, lower Bavaria (Germany) Q. robur GWORO977-09 Lower Saxony (Germany) – GWORO967-09 Lower Saxony (Germany) – GWORO954-09 Lower Saxony (Germany) – Operophtera fagata Adult GBLAC168-13 Saarland (Germany) – GBLAC993-13 Saxony (Germany) – GBLAF596-14 Brandenburg, Barnim (Germany) – GWOTD330-12 Calabria (Italy) – GWOTD332-12 Calabria (Italy) – GWORB791-07 Bavaria, south (Germany) – GBLAC167-13 Saarland (Germany) – Larva GWORO972-09 Lower Saxony (Germany) –

(−) Absence of information about host plant in database

2 2 3Results plants (χ 2 = 13.03, p < 0.001) and years (χ 2 =13.78, p < 0.001). The interaction term was also significant 2 3.1 Tree density and budburst variation (χ 4 = 18.89, p < 0.001). Budburst of Q. afares and Q. canariensis began in late March, but budburst of Q. canariensis was the most abundant species in the stud- Q. suber occurred about 3 weeks later (Fig. 1). In 2010 ied region with an average density of 700 trees/ha follow- and 2011, 50% of Q. canariensis budburst occurred 1 week ed by Q. suber and Q. afares with an average density of before Q. afares. In 2012, budburst of Q. canariensis be- 175 trees/ha and 60 trees/ha, respectively. The Julian day gan a week after Q. afares. For all years, Q. suber when 50% of budburst occurred varied between host budburst was 3 to 4 weeks later (Fig. 1).

Fig. 1 Evolution of budburst of Q. canariensis, Q. afares and Q. suber in 2010, 2011, and 2012 Annals of Forest Science (2017) 74: 3 Page 5 of 8 3

3.2 Larval density on host plants 3.5 Molecular data and patterns

For 6 weeks, a total of 1057, 1011, and 355 larvae were col- A total of 15 barcode sequences of Tunisian O. brumata were lected in 2010, 2011, and 2012, respectively. The average sequenced to the full barcode region of 658 bp (BOLD, pro- number of larvae per branch varied between host species ject INRGR http://www.barcodinglife.com/Global 2 2 (χ 2 = 883.12, p < 0.001) and years (χ 2 = 115.50, Geometridae/Lepidoptera of Tunisia-cork) represented in p < 0.001); the interaction term was also significant Fig. 4. No genetic differences were found between the speci- 2 (χ 4 = 112.50, p < 0.001). For all weeks, mean larval numbers mens feeding on the various host plants. All 15 Tunisian DNA were higher on Q. canariensis than on Q. afares and Q. suber barcodes belong to exactly the same haplotype and perfectly (Table 2). For all host species, larval density was highest in match the haplotype of the populations examined from a April (Table 2). Tunisian forest (Mzara) next to Ain Zena, as well as those from Spain, Italy, and Germany (see Fig. 4). On the BOLD database, there are additional data of the same haplotype from 3.3 Larval development and mortality Morocco, France, the Netherlands, Austria, UK, and Canada. Populations with slightly diverging haplotypes (diverging by In the laboratory experiment, total larval development time 1–2 basepairs only) have been barcoded from northern from the 1st instar to the 5th was shorter on Q. canariensis Germany, Finland, and UK The genetic distance from O. (36.5±0.3days)thanonQ. afares (40±0.3days)and fagata,chosenasoutgroup,was7.0%(Fig.4). Q. suber (46.2 ± 0.3 days). The host species had a significant effect on larval development (F(2, 206) = 245.84, p<0.001). For each instar larva, the development was faster on Q. canariensis than on Q. afares and Q. suber (Fig. 2). 4Discussion Death cause was in most cases unknown. Mortality was higher on Q. afares for the 1st and 2nd instar larvae and great- The winter moth is a polyphagous insect that takes advantage er on Q. suber for the 4th and 5th instar larvae (Fig. 3). The of many different host species, when available; oak is usually the primary host with a high density and greater defoliation host species had a significant effect on the proportion of 3rd 2 2 (O’Donnell 2015). Larval density was high in 2010 and 2011 (χ 1 = 11.34, p <0.001)and4th(χ 2 =7.3,p <0.05)instar dead larvae. on all the oak species considered, especially on Q. canariensis. Then in 2012, the density fell sharply (Table 2). In the field, Q. canariensis trees had a higher den- 3.4 Pupal weight sity on average than the other two oaks. It was the most infested host species. Cunningham et al. (2001)showedthat The pupal weight varied significantly among species, in the when the abundance of one host species is high, the probabil- laboratory trials (F(2, 50) =38.12,p<0.001) and in the field ity that the insect will land on this species is greater. (F(2, 73) =13.96,p<0.001). It was higher for Q. canariensis Furthermore, phenological differences among and within host with an average of 27.2 ± 0.5 mg and 28.8 ± 0.5 mg, from species are very important factors which affect host use and laboratory feeding trials and the field, respectively, than adaptations of O. brumata. Many works have shown that there Q. afares at 24.4 ± 0.6 mg and 26.5 ± 0.5 mg and Q. suber is a large annual variation of budburst of oak (Wint 1983; at 20.4 ± 0.6 mg and 23.8 ± 0.9 mg. Fraval 1984;DuMerle1988; Van Dongen et al. 1997;Pinto

Table 2 Mean number of larvae on each host species per week and per year

Year Host plant W1 W2 W3 W4 W5 W6 Mean number of larvae

2010 Q. canariensis 3.1 ± 1.4 7.5 ± 2.7 14.7 ± 7.2 23.2 ± 7 21.2 ± 6.2 15.5 ± 3.6 12 ± 2 Q. afares 0.4 ± 0.3 3.5 ± 1.3 4.6 ± 1.6 2.7 ± 0.8 2.2 ± 1 1.9 ± 0.6 2.7 ± 0.4 Q. suber 0 0 0 1.9 ± 0.9 1.9 ± 0.8 1.4 ± 0.8 1.7 ± 0.6 2011 Q. canariensis 1.2 ± 0.4 9.4 ± 2 24.2 ± 5.7 20.8 ± 6.3 16.8 ± 4.4 4.9 ± 2.1 12.9 ± 2 Q. afares 0 3.6 ± 0.9 5.9 ± 1.2 3.7 ± 0.8 3.1 ± 1 0.8 ± 0.5 2.8 ± 0.4 Q. suber 0 0 0 0 2.5 ± 1.4 4.2 ± 1.5 1.13 ± 0.3 2012 Q. canariensis 3.5 ± 1.2 2.4 ± 1 6.3 ± 2.6 4 ± 1.3 2.3 ± 1 2.7 ± 1 3.5 ± 0.6 Q. afares 1.7 ± 0.6 2.4 ± 0.7 3.8 ± 1.3 3.3 ± 1.2 1.1 ± .5 0.8 ± 0.4 2.2 ± 0.4 Q. suber 0 0 0 0 0.8 ± 0.5 0.4 ± 0.2 0.2 ± 0.1

W1 3rd week of March, W2 4th week of March, W3 1st week of April, W4 2nd week of April, W5 3rd week of April, W6 4th week of April 3 Page 6 of 8 Annals of Forest Science (2017) 74: 3

Fig. 2 Development time of O. brumata in days (±SE) from first (L1) to fifth (L5) larval instar

et al. 2011). Budburst timing did not differ between of the host plant on larval performance (Hunter 1992; Q. canariensis and Q. afares whereas that of Q. suber oc- Tikkanen and Lyytikainen-Saarenmaa 2002). O. brumata per- curred about 3 weeks later than the other two species formance varies among these three Mediterranean oak species (Fig. 1). In a given region with several host species, which would offer different food quality. Budburst of O. brumata will feed mainly on the host which budburst phe- Q. canariensis and Q. afares occurred at the same time. The nology coincides with its larval hatching. Once larvae shortest larval development time was recorded for larvae feed- emerged, they needed young leaves available on deciduous ing on Q. canariensis which was the most infested host, while to semi-evergreen Quercus species: Q. canariensis and the longest development time was recorded for those feeding Q. afares. On the evergreen oak Q. suber, larvae found only on Q. suber (Fig. 2). O’Donnell (2015)showedthatlarvaeof old leaves, which explain the absence of young larvae on this O. brumata fed on Q. rubra L. and Malus domestica Borkh species (Table 2). This suggests that the phenology of had faster development and lower mortality than larvae fed on Q. canariensis allows a high colonization by O. brumata. Acer rubrum L., Betula papyrifera Marshall, Q. canariensis probably offers more chances for O. brumata pensylvanica L.f., Vaccinium angustifolium Aiton, and to survive because of early leafing and intense budburst which V. corymbosum L. Similarly, Ruuhola et al. (2001)alsoob- exceeds 50% at the peak density of the insect (Fig. 1, Table 2). served that larval growth of the winter moth on Salix Hunter (1992) found that Quercus robur L. trees that leaf out phylicifolia L. was significantly faster than on Salix pentandra early have the highest density of . L. and on Salix myrsinifolia Salisb. Larval growth is clearly The suitability of host species for larva development and better on certain hosts than others, depending on host nutri- survival of polyphagous insects differs from one host plant tional quality and resistance mechanisms (Kirsten and Topp species to another. Generalist Lepidoptera species can vary 1991). The ability of generalist Lepidoptera to complete larval greatly in their growth efficiency on different natural host development, even at the cost of increased development times plants (Feeny 1970). Many studies have focused on the effect (Wint 1983), is the most important feature of their polypha- gous habit (Warrington 1985). Results of the pupal weight of O. brumata suggest that food quality may have influenced pupal mass. Tikkanen (2000) demonstrated that pupal weight of O. brumata larvae reared on Q. robur varied between 16 and 42.2 mg. In the data obtained here, pupae of larvae reared on Q. canariensis were heavier than those reared on Q. afares and Q. suber. According to the slow-growth-high-mortality hypothesis (Clancy and Price 1987), the extended feeding period makes insect larvae more susceptible to attacks by predators, parasit- oids, and pathogens, resulting in a higher mortality in natural environments (Häggström and Larsson 1995). Host species Fig. 3 Mortality rate (±SE) of each larval stage (first (L1) to fifth (L5) also affects mortality of O. brumata larvae (Wint 1983). larval instar) collected from the field Mortality of 4th and 5th instar larvae was greater on Annals of Forest Science (2017) 74: 3 Page 7 of 8 3

Fig. 4 Maximum Likelihood (ML) Tree including 21 Operophtera bootstrap method, 500 replicates, Tamura-Nei model, complete specimens from the western Palearctic (3 from Tunisia), constructed deletion, bootstrap values indicated when >50% with MEGA6 and MEGA7 (Tamura et al. 2013; Kumar et al. 2016),

Q. suber than on Q. canariensis and Q. afares (Fig. 3). 5 Conclusion Caterpillars would then experience higher rates on Q. suber (which should be tested in further experiments). We concluded that the deciduous Q. canariensis was Molecular data and patterns show no genetic difference more susceptible to infestation by the winter moth than between the Tunisian populations and the European “main other species due to its high density in the field, early pool” of that species belonging all to one and the same leafing, and best food quality offered to this insect. haplotype (Fig. 4). This result (1) agrees with the absence There are significant differences in the performance of of host specialization (potentially leading to complex hap- larval development on the different hosts. These differ- lotype diversification) and the polyphagous life history of ences are reflected in the various components of indi- this species; and (2) clearly supports the hypothesis of a vidual fitness, such as larval development time, larval recent expansion of the distribution area instead of an mortality, and pupal weight. overlooked occurrence of a long-term isolated population Population abundance of the winter moth observed after in North Africa. First instar larvae of O. brumata are 2011 was not sufficient to damage trees. However, this insect abseilingonsilkythreadsenablingtolong-distancedis- was observed in other cork oak forests (Ain El Baya in 2013 persal by wind (Hausmann and Viidalepp, 2012). and El Jouza in 2015), suggesting further expansion. Winter Moreover, the species has been recorded in Morocco moth populations should be managed through constant mon- (Hausmann and Viidalepp 2012), supposedly colonized itoring for early detection of outbreaks. from Spain. The latter country, therefore, has to be regarded as a potential origin of the Tunisian populations as well as Italy. The hypothesis of a recent colonization Acknowledgments Thanks to Mabrouk GRAMI (CFAR, Tunisia) and forest technicians for their valuable help in the field. The due to an anthropogenous transport of immature stages genetic analyses have received considerable support from Paul over long distances with tree seedlings is not excluded, D. N. HEBERT and the Biodiversity Institute of Ontario (BIO) but is questioned by the fact that Tunisia has not seen and the Canadian Centre for DNA Barcoding (CCDB University any colonization from one of the many other Central of Guelph). The data management and analysis system BOLD was provided by Sujeevan RATNASINGHAM. We wish to thank European defoliating moth species (Hausmann and Christie NIELSEN CHAAR and Emna DARGHOUTHI for their Viidalepp 2012). assistance with the language editing. 3 Page 8 of 8 Annals of Forest Science (2017) 74: 3

Compliance with ethical standards Kumar S, Stecher G, Tamura K (2016) MEGA7: molecular evolutionary genetics analysis version 7.0 for bigger datasets. Mol Biol Evol 33: – Funding Funding was provided to achieve the presented results and was 1870 1874. doi:10.1093/molbev/msw074 given from the National Institute for Research in Rural Engineering, Water Kúti ZS, Hirka A, Hufnagel L, Ladanyl M (2011) A population dynam- and Forest (INRGREF) and Genome Canada (Ontario Genomics Institute) ical model of Operophtera brumata L. extended by climatic factors. in the framework of the iBOL program, WG 1.9. Appl Ecol Env Res 9:433–447. doi:10.15666/aeer/0904_433447 Mannai Y, Ezzine O, Nouira S, Ben Jamâa ML (2015) First report of the winter moth Operophtera brumata on Quercus canariensis and Q. afares in north west of Tunisia. TJPP 10:69–74 References Mhamdi S, Brendel O, Montpied P, Ghouil-Amimi H, Hasnaoui I, Dreyer E (2013) Leaf morphology displays no detectable spatial organiza- tion in the relict Quercus afares Pomel compared to the co-occurring Clancy KM, Price PW (1987) Rapid herbivore growth enhances enemy parental species Q. canariensis Willd and Q. suber L. Ann For Sci – attack: sublethal plant defenses remain a paradox. Ecology 68:733 70:675–684. doi:10.1007/s13595-013-0312-x 737. doi:10.2307/1938479 Mir C, Toumi L, Jarne P, Sarda V, Di Giusto F, Lumaret R (2006) Endemic Cunningham JP, West SA, Zalucki MP (2001) Host selection in phytoph- North African Quercus afares Pomel originates from hybridisation agous insects: a new explanation for learning in adults. Oikos 95: between two genetically very distant oak species (Q. suber L. and Q. – 537 543. doi:10.1034/j.1600-0706.2001.950319.x canariensis Willd.): evidence from nuclear and cytoplasmic markers. Du Merle P (1988) Phenological resistance of oaks to the green leafroller, Heredity 96:175–184. doi:10.1038/sj.hdy.6800782 Tortrix viridiana (Lepidoptera: Tortricidae). In: Mattson WJ, O’Donnell K M (2015) The relationship between the winter moth Levieux J, Bernard-Dagan C (eds) Mechanisms of woody plant (Operophtera brumata) and its host plants in Coastal . defenses against insects. Springer, Berlin Heidelberg New York, Electronic Theses and Dissertations. University of Maine pp 215–226 Pinto CA, Henriques MO, Figueiredo JP,David JS, Abreu FG, Pereira JS, Feeny P (1970) Seasonal changes in oak-leaf tannins and nutrients as a Correia I, David TS (2011) Phenology and growth dynamics in cause of spring feeding by winter moth caterpillars. Ecology 51: Mediterranean evergreen oaks: effects of environmental conditions 565–581. doi:10.2307/1934037 and water relations. For Ecol Manag 262:500–508. doi:10.1016/j. Foster JR, Townsend PA, Mladenoff DJ (2013) Mapping asynchrony foreco.2011.04.018 between gypsy moth egg-hatch and forest leaf-out: putting the phe- Ratnasingham S, Hebert PDN (2007) BOLD: the barcode of life data nological window hypothesis in a spatial context. For Ecol Manag system. Mol Ecol Notes 7:355–364. doi:10.1111/j.1471- 287:67–76. doi:10.1016/j.foreco.2012.09.006 8286.2007.01678.x Fraval A (1984) Influence de la qualité et de la quantité de l’alimentation Ruuhola T, Tikkanen OP, Tahvanainen J (2001) Differences in host use sur les fluctuations des populations de Lymantria dispar L. (Lep. efficiency of larvae of a generalist moth, Operophtera brumata of Lymantriidae) en forêt de la Mamora (Maroc). Agronomie 4:819– three chemically divergent Salix species. J Chem Ecol 27:1595– 828. doi:10.1051/agro:19840903 1615. doi:10.1023/A:1010458208335 Häggström H, Larsson S (1995) Slow larval growth on a suboptimal Tamura K, Stecher G, Peterson D, Filipski A, Kumar S (2013) MEGA6: results in high predation mortality in the leaf beetle molecular evolutionary genetics analysis version 6.0. Mol Biol Evol Galerucella lineola. Oecologia 104:308–315. doi:10.1007 30:2725–2729. doi:10.1093/molbev/mst197 /BF00328366 Tikkanen OP (2000) Adaptation of a generalist moth, Operophtera Hausmann A, Viidalepp J (2012) The geometrid of Europe, vol 3. brumata, to host plants. Dissertation, University of Joensuu Apollo Books, Vester Skerninge Tikkanen OP, Julkunen-Tiitto R (2003) Phenological variation as protec- tion against defoliating insect: the case of Quercus robur and Hebert PDN, Ratnasingham S, De Waard JR (2003) Barcoding animal – life: cytochrome c oxidase subunit 1 di-vergences among closely Operophtera brumata. Oecologia 139:244 251. doi:10.1007 related species. Proc R Soc Lond B Biol Sci 270:S96–S99. /s00442-003-1267-7 doi:10.1098/rsbl.2003.0025 Tikkanen OP, Lyytikainen-Saarenmaa P (2002) Adaptation of a generalist moth, Operophtera brumata, to variable budburst phenology of host Holliday NJ (1977) Population ecology of winter moth (Operophtera plants. Entomol Exp Appl 103:123–133. doi:10.1046/j.1570- brumata) on in relation to larval dispersal and time of 7458.2002.00966.x budburst. J Appl Ecol 14:803–813. doi:10.2307/2402812 Van Asch M, Julkunen-Tiito R, Visser ME (2010) Maternal effects in an Horgan FG (1993) Factors affecting the mortality of winter moth in the insect herbivore as a mechanism to adapt to host plant phenology. lower mainland of British Columbia. Dissertation, University of Funct Ecol 24:1103–1109. doi:10.1111/j.1365-2435.2010.01734.x British Columbia, Vancouver Van Dongen S, Backeljau T, Matthysen E, Dhondt AA (1997) Hunter MD (1990) Differential susceptibility to variable plant phenology Synchronization of hatching date with budburst of individual host and its role in competition between two insect herbivores on oak. Ecol trees (Quercus robur) in the winter moth (Operophtera brumata) – Entomol 15:401 408. doi:10.1111/j.1365-2311.1990.tb00823.x and its fitness consequences. J Anim Ecol 66:113–121. Hunter MD (1992) A variable insect-plant interaction: the relationship doi:10.2307/5969 between tree budburst phenology and population levels of insect Warrington S (1985) Consumption rates and utilization efficiencies of herbivores among trees. Ecol Entomol 16:91–95. doi:10.1111 four species of polyphagous Lepidoptera feeding on sycamore /j.1365-2311.1992.tb01046.x leaves. Oecologia 67:460–463. doi:10.1007/BF00790015 Ivanova V, Dewaard JR, Hebert PDN (2006) An inexpensive, Winstad OPL, Hagen SB, Jepsen SB, Kapari L, Schott T, Ims RA (2011) automation friendly protocol for recovering high quality Phenological diversity in the interactions between winter moth DNA.MolEcolNotes:6:998–61002. doi:10.1111/j.1471- (Operophtera brumata) larvae and parasitoid wasps in sub-arctic x8286.2006.01428.x mountain forest. Bull Entomol Res 101:705–714. Kirsten K, Topp W (1991) Acceptance of willow-species for the devel- doi:10.1017/S0007485311000277 opment of the winter moth, Operophtera brumata (Lep., Wint W (1983) The role of alternative hostplant species in the life of a Geometridae). J Appl Entomol 111:457–468. doi:10.1111/j.1439- polyphagous moth, Operophtera brumata (Lepidoptera: 0418.1991.tb00348.x Geometridae). J Anim Ecol 52:439–450. doi:10.2307/4564